Not that the plans aren’t valid…they are (I built the same physics optimization model 8 years ago; search old posts on here). But, from playing around with these models it all ends up very simple. I loaded the Oceanside course in some detail, and set up a spreadsheet that computed power/speed for every road segment, including inputs for CdA, Crr, weight, etc. I then set it up to optimize for lowest time, while giving the model freedom to vary the power on each road segment, subject to a global NP limit. It worked…I rode as the model advised and my bike split was within a minute of predicted.
But…there is NO NEED to try to micro-manage the ride. It’s way simpler than that and you don’t need a custom model. I did dozens of hypothetical course models and quickly realized that the optimal strategy anywhere comes down to simple heuristics.
First, realize that only *you *can decide your NP “budget.” You gotta figure that out. Let’s say you choose 200 watts NP for a half IM.
Now…how do you ride each individual course segment such that (1) your NP does not exceed 200w; and (2) in the shortest amount of time?
It’s an optimization problem, and you can set it up in Excel (or use BBS.com).
And the answer is…ride very nearly exactly constant the whole time. Ride slightly harder on hills and *slightly *easier on downhills. Ignore wind – ride the same since the *optimal *power variation upwind vs. downwind is so small that you can’t ride that accurately anyway.
Lift power maybe 5% on hills, and drop 5% on descents (until you run out of gears). On really steep ups/downs, raise that to 10% if it doesn’t go on for too long.
On flats, sit on your power goal.
For race-prep training rides, see if you can get your ride NP/AP consistently under 1.03. For all but the hilliest rides, the lower that ratio, the better. Even then, anything over 1.05 is losing time.
Don’t overthink it. Ride steady with slight lifts on uphills, and slight drops on downhills. Ignore wind.
